use Perl

The June meeting of SF.pm will take place at 7 pm PST on June 23rd at Six Apart World Headquarters.

Jeff Thalhammer returns to speak on "Continuous Integration of Perl Code". The discussion will cover:

        * What is Continuous Integration (CI)?
        * A survey of available CI frameworks.
        * Tips and tricks for using CI with Perl.
        * Show-and-tell with an actual CI system.

Announcement posted via App::PM::Announce

RSVP at Meetup - http://www.meetup.com/San-Francisco-Perl-Mongers/calendar/10058948/

Reposting the announcement by Philippe M. Chiasson (Gozer) from the mod_perl users list

http://perl.apache.org/dist/

It's finally here! After quite a few release candidates, I am happy to share with you all this new mod_perl release. Enjoy!

Changes since 1.30:

• Fix static APACI build on Mac OS X. [Gozer]
• Fix XSS vulnerability in Apache::Status reported by Richard J. Brain, CVE-2009-0796 [Fred Moyer]
• On Win32, mod_perl.h needs to include before the perl headers, at least when built with USE_ITHREADS [Steve Hay]
• Win32 needs PERL_SYS_INIT/PERL_SYS_TERM calls when built with USE_ITHREADS [sic--that's different to USE_THREADS]. In fact, they ought to be always called if they are defined [Steve Hay]
• Fix potential segfault when the environment contains NULL values [Mike Schilli]
• Fix static APACI build against newer apache-1.3.38+ [Gozer]
• Fixed modules/regex.t test 4 on Win32 [Steve Hay]
• Avoid possible segfault when PerlFreshRestart is On. [Michael Rendell ]
• Prevent segfault when running with perl >= 5.9.3 [Steve Hay]
• Fix shared libary extensions on Win32 to be .dll not .so [Nikolay Ananiev ]
• Patch to mod_perl.dsp to remove /D _WINSOCK2API_ on Win32 for perl >= 5.8.6 [Steve Hay]

Update - the Perlmonks node.

This is my latest attempt at a Makefile.PL section that invokes Apache::Bootstrap or exits with 255. The intent is that if Apache::Bootstrap is not installed, the Makefile.PL dies with exit code 255, the same failure code returned to the shell as if 'use 5.8.8;' failed in Makefile.PL. I've been looking around for a list where the MakeMaker experts hang out, please clue me in if you know of such a place.

#!/usr/bin/perl

use strict;
use warnings;

################################################################### ######
# minimum version of A::B required
my $ab = 0.06;

eval { require Apache::Bootstrap };

($@ or ($Apache::Bootstrap::VERSION < $ab)) &&
    do { $! = 255; die("Apache::Bootstrap $ab required"); };

# make sure we have at least one minimum version of mod_perl

my $bs = eval { Apache::Bootstrap->new({
                      mod_perl2 => 1.99022,
                      mod_perl  => 1.30 }) };

do { $! = 255; die($@); } if $@;

######################################################################### #

# use $bs to determine whether ExtUtils::MakeMaker or ModPerl::Build::MM should be used

I fixed a four year old bug in Apache::Test tonight, and closed another bug that is no longer relevant. There were 9 open bugs total, now there are only 7 more to fix.

==> Auto-install the 3 mandatory module(s) form CPAN? [y] n

==> The module(s) are mandatory! Really skip? [n] y

If I type 'n', do I really need to be reminded again??

I've been working on a system that requires exceptionally high uptime. As such, making major releases takes a couple weeks of testing and preparation. I usually end up doing these releases at 2 am; they are fully automated and things almost always go off without a hitch.

It is somewhat anticlimactic though - there is a 30 second service interruption while perlbal and mod_perl shut down, the postgresql database is upgraded in place, and then perlbal and mod_perl start up again. So all of my work is done upfront; reminds me of the scene from the movie "Lost in Space" where the Jupiter One is ready to launch, and the pilot says "And the monkey pushes the button."

I'm still tracking down the details, but it appears that recent versions of the linux kernel headers fail to include /usr/include/asm/page.h. The lack of that file cause my perl 5.8.8 build to fail, SysV.xs (or something similar) was including that header file. For now I'm going with the 2.6.17 headers and the 2.6.22 kernel (installed via Gentoo).

Dear LazyPerl,

I had responsibility for a ROR (Ruby on Rails) forum called Beast land squarely in my lap today. Mission critical resource, of course. The box it was being hosted on was upgraded to ruby 1.87, which has some non-backwards compatible upgrades.

Needless to say, the Beast application was not happy, and users cannot post to the forum. Beast 1, Phred 0. The Beast website is also offline. Beast 2, Phred 0.

Do you know a Ruby contractor hacker that can help me even the score? Email fred at redhotpenguin dot com.